Skip to main content
BenchLM
Data

Grok 4.5 vs Kimi K2.6

Updated November 21, 2026. Rank says Grok 4.5 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.

Share or export
Share on XLinkedInAPI/MCP

Decision reading

Grok 4.5 has the higher public score estimate, 60.4 versus 54.3, but the 90% score intervals overlap. Treat that as a lead, not a settled winner. 10 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A
xAI logo

xAI

60.4/100

Supported · Public rank #53

90% interval 56.3–64.5

Model B
Moonshot AI logo

Moonshot AI

54.3/100

Supported · Public rank #76

90% interval 50.3–58.4

API price per 1M tokens

Grok 4.5: $2 input / $6 output

Kimi K2.6: $0.95 input / $4 output

Context window

Grok 4.5: 500K

Kimi K2.6: 256K

Released

Grok 4.5: July 8, 2026

Kimi K2.6: April 20, 2026

Shared results
10
Grok 4.5 only
12
Kimi K2.6 only
27
Like-for-like categories
5 / 8
Supported: Grok 4.5 and Kimi K2.6How the comparison works

Which one for your work

Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.

  • Coding work

    Code generation, repair, and software-engineering tasks

    Grok 4.5

    Grok 4.5 has the higher public coding point estimate, 60.5 to 52.9, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

    Confidence: limited
  • Agentic work

    Tool use, computer use, and multi-step task completion

    Grok 4.5

    Grok 4.5 has the higher public agentic point estimate, 57.8 to 54.4, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

    Confidence: limited
  • Long documents

    Prompts that approach the documented context limit

    Grok 4.5

    Grok 4.5 has the larger documented context window.

    Confidence: documented
Show secondary and unsupported calls
  • Chat turn cost

    1K fresh input + 500 output tokens

    Kimi K2.6

    Kimi K2.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates
  • Cache-heavy agent loop cost

    200K cached + 20K fresh input + 10K output tokens

    Kimi K2.6

    Kimi K2.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates
  • Repository review cost

    50K fresh input + 3K output tokens

    Kimi K2.6

    Kimi K2.6 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates

Which one for a specific job

Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

60.5Grok 4.552.9Kimi K2.6

Like-for-like · BenchAlign v6.0

Grok 4.5 has the higher coding point estimate. Conditional score ranges do not establish rank confidence.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Same basis rules as the category table below

Or describe your own job

The LLM Selector reads what you need a model for and returns a shortlist, ranked in the same order the job pages publish.

Describe the task and constraints. This selector recommends models; it cannot carry out the task or access your files and accounts.

We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.

Shape of the matched evidence

Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.

Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.

Bars run 0–100 on each benchmark’s normalized display scale

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v6.0 lane in every category, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence. Directional and not-comparable rows remain visible, but they never receive a winner.

Agentic

Like-for-like
Grok 4.5
57.8
Supported · #65/182
Kimi K2.6
54.4
Supported · #80/182
Basis
BenchAlign v6.0 lane · 6 vs 12 public rows
Reading
Grok 4.5 leads · intervals overlap

Coding

Like-for-like
Grok 4.5
60.5
Supported · #54/180
Kimi K2.6
52.9
Supported · #71/180
Basis
BenchAlign v6.0 lane · 9 vs 10 public rows
Reading
Grok 4.5 leads · intervals overlap

Reasoning

Like-for-like
Grok 4.5
59.9
Supported · #48/161
Kimi K2.6
51.5
Supported · #73/161
Basis
BenchAlign v6.0 lane · 3 vs 0 public rows
Reading
Grok 4.5 leads · intervals overlap

Instruction following

Like-for-like
Grok 4.5
57.7
Supported · #40/143
Kimi K2.6
52.4
Supported · #60/143
Basis
BenchAlign v6.0 lane · 0 vs 0 public rows
Reading
Grok 4.5 leads · intervals overlap

Math

Like-for-like
Grok 4.5
57.5
Supported · #42/125
Kimi K2.6
53.5
Supported · #48/125
Basis
BenchAlign v6.0 lane · 2 vs 5 public rows
Reading
Grok 4.5 leads · intervals overlap

Multimodal

Directional only
Grok 4.5
62.9
Supported · #43/119
Kimi K2.6
48.3
Estimated · #77/119
Basis
BenchAlign v6.0 lane · 0 vs 5 public rows
Reading
Directional only

Knowledge

Directional only
Grok 4.5
61.3
Supported · #39/158
Kimi K2.6
53.5
Estimated · #70/158
Basis
BenchAlign v6.0 lane · 2 vs 5 public rows
Reading
Directional only

Multilingual

Directional only
Grok 4.5
59.0
Estimated · #27/88
Kimi K2.6
56.9
Estimated · #32/88
Basis
BenchAlign v6.0 lane · 0 vs 0 public rows
Reading
Directional only

Ranks count the models scored in each category’s BenchAlign v6.0 lane, so denominators differ by category. Like-for-like rows within 0.5 points read as a practical tie.

Supported evidence per lane · bars run 0–100Methodology

What each workload costs

Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.

Chat turn

1K fresh input + 500 output tokens

Grok 4.5
$0.005
Fits in one request
Kimi K2.6
$0.00295
Fits in one request

Kimi K2.6 has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

Grok 4.5
$0.118
Fits in one request
Kimi K2.6
$0.0595
Fits in one request

Kimi K2.6 has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

Grok 4.5
$0.16
Fits in one request
Kimi K2.6
$0.091
Fits in one request

Kimi K2.6 has the lower modeled cost

Costs use the listed standard API rates.

Your workload

Enter a month of tokens to price it at both models’ listed API rates.

Grok 4.5
$32.00
per month
Kimi K2.6
$17.50
per month

Standard listed rates, no batch or volume discounts.

Cached input falls back to the list input rate only where a cached rate is unpublished

Price against score, among all priced models

Overall score against output price per 1M tokens, one model per family. Pricing as of November 21, 2026.

Grok 4.5Kimi K2.6

20406080100$0.10$0.50$1$5$10$50$100Output price, $ per 1M tokens (log)Overall score
  • 12 models score at least as high as Grok 4.5 for the same price or less. The nearest is Grok 4.6 (score 65.2, also $6.00 per 1M output tokens).
  • 12 cheaper models score at least as high as Kimi K2.6. The nearest is DeepSeek V4 Pro (score 61.6, $3.96 per 1M output tokens).

Gray dots are the other priced models. The dark step line is the efficiency frontier: for a model under it, another model costs no more and scores at least as high.

See Kimi K2.6 on the full price vs performance chart

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

Context window

Maximum documented context; output-token limits may be lower.

Grok 4.5

500K

Kimi K2.6

256K

API model ID

Grok 4.5

Not sourced

Kimi K2.6

Not sourced

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

Grok 4.5

$0.3 per 1M cached input tokens

Kimi K2.6

$0.16 per 1M cached input tokens

Provider availability

Grok 4.5

Not sourced

Kimi K2.6

Not sourced

Reasoning profile

Grok 4.5

Reasoning

Kimi K2.6

Reasoning

Weight access

Grok 4.5

Proprietary

Kimi K2.6

Open Weight

License

Grok 4.5

Proprietary

Kimi K2.6

Open Weight

Release date

Grok 4.5

2026-07-08

Kimi K2.6

2026-04-20

If you already use one of these models

Deployment change
The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
Quality signal
Grok 4.5 has the higher public score estimate, 60.4 versus 54.3, but the 90% score intervals overlap.
Workload cost
Repository review: $0.118 vs $0.0595. Cache-heavy agent loop: $0.16 vs $0.091.
Context tradeoff
Grok 4.5 has the larger documented window (500K).
Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, Grok 4.5 or Kimi K2.6?

Grok 4.5 has the higher public score estimate, 60.4 versus 54.3, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.

Which is better for coding, Grok 4.5 or Kimi K2.6?

Grok 4.5 has the higher public coding point estimate, 60.5 to 52.9, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which is better for agentic tasks, Grok 4.5 or Kimi K2.6?

Grok 4.5 has the higher public agentic tasks point estimate, 57.8 to 54.4, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which costs less, Grok 4.5 or Kimi K2.6?

For the stated presets, chat costs $0.005 on Grok 4.5 and $0.00295 on Kimi K2.6; repository review costs $0.118 and $0.0595; the cache-heavy agent loop costs $0.16 and $0.091. Costs use the listed standard API rates.

Which has the larger context window, Grok 4.5 or Kimi K2.6?

Grok 4.5 has the larger documented context window: 500K, compared with 256K.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Grok 4.5
API / mo$6,000
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K2.6
API / mo$3,713
Self-host / mo$27,962
Break-evenNot reached at 70% load
Model the full break-even

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence49 rows

Agentic

  • Grok 4.547.24%
    Kimi K2.646.99%

    Grok 4.5 leads this result

  • Terminal-Bench 4.1

    Grok 4.512.40%
    Source
    Kimi K2.6—

    Not directly comparable

  • Terminal-Bench 3.0

    Grok 4.515.7%
    Source
    Kimi K2.6—

    Not directly comparable

  • Terminal-Bench 2.1

    Grok 4.583.3%
    Source
    Kimi K2.6—

    Not directly comparable

  • deepSwe

    Grok 4.553%
    Source
    Kimi K2.6—

    Not directly comparable

  • Terminal-Bench 2.1 (Vals)

    Grok 4.567.8%
    Source
    Kimi K2.653.6%
    Source

    Grok 4.5 leads this result

  • Terminal-Bench 2.0

    Grok 4.5—
    Kimi K2.666.7%
    Source

    Not directly comparable

  • BrowseComp

    Grok 4.5—
    Kimi K2.683.2%
    Source

    Not directly comparable

  • OSWorld-Verified

    Grok 4.5—
    Kimi K2.673.1%
    Source

    Not directly comparable

  • Toolathlon

    Grok 4.5—
    Kimi K2.650%
    Source

    Not directly comparable

  • MCP Atlas

    Grok 4.5—
    Kimi K2.655.9%
    Source

    Not directly comparable

  • Claw-Eval

    Grok 4.5—
    Kimi K2.662.3%
    Source

    Not directly comparable

  • DeepSearchQA

    Grok 4.5—
    Kimi K2.692.5%
    Source

    Not directly comparable

  • WideResearch

    Grok 4.5—
    Kimi K2.680.8%
    Source

    Not directly comparable

  • ResearchClawBench

    Grok 4.5—
    Kimi K2.618.0%
    Source

    Not directly comparable

  • OSWorld 2.0

    Grok 4.5—
    Kimi K2.64.6%
    Source

    Not directly comparable

Coding

  • SWE-bench Pro

    Grok 4.564.7%
    Source
    Kimi K2.658.6%
    Source

    Grok 4.5 leads this result

  • SWE Multilingual

    Grok 4.578%
    Source
    Kimi K2.676.7%
    Source

    Grok 4.5 leads this result

  • Terminal-Bench 2.1

    Grok 4.583.3%
    Source
    Kimi K2.6—

    Not directly comparable

  • CursorBench 3.2

    Grok 4.566.7%
    Source
    Kimi K2.6—

    Not directly comparable

  • VulcanBench v3

    Grok 4.589.9%
    Source
    Kimi K2.6—

    Not directly comparable

  • LiveCodeBench (Vals)

    Grok 4.587.4%
    Source
    Kimi K2.686.8%
    Source

    Grok 4.5 leads this result

  • SWE-bench (Vals)

    Grok 4.586.6%
    Source
    Kimi K2.676.2%
    Source

    Grok 4.5 leads this result

  • PostTrainBench v1.1

    Grok 4.523.4%
    Source
    Kimi K2.6—

    Not directly comparable

  • PostTrainBench

    Grok 4.522.6%
    Source
    Kimi K2.6—

    Not directly comparable

  • SWE-bench Verified

    Grok 4.5—
    Kimi K2.680.2%
    Source

    Not directly comparable

  • LiveCodeBench v6

    Grok 4.5—
    Kimi K2.689.6%
    Source

    Not directly comparable

  • SciCode

    Grok 4.5—
    Kimi K2.652.2%
    Source

    Not directly comparable

  • Terminal-Bench 2.0

    Grok 4.5—
    Kimi K2.666.7%
    Source

    Not directly comparable

  • Vibe Code Bench

    Grok 4.5—
    Kimi K2.637.89%
    Source

    Not directly comparable

  • cursorBench31

    Grok 4.5—
    Kimi K2.647.6%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    Grok 4.552.6%
    Source
    Kimi K2.6—

    Not directly comparable

  • ARC-AGI-3

    Grok 4.50.3%
    Source
    Kimi K2.6—

    Not directly comparable

  • ARC-AGI-1

    Grok 4.585.67%
    Source
    Kimi K2.6—

    Not directly comparable

Multimodal

  • MMMU-Pro

    Grok 4.5—
    Kimi K2.679.4%
    Source

    Not directly comparable

  • MMMU-Pro w/ Python

    Grok 4.5—
    Kimi K2.680.1%
    Source

    Not directly comparable

  • CharXiv

    Grok 4.5—
    Kimi K2.680.4%
    Source

    Not directly comparable

  • MathVision

    Grok 4.5—
    Kimi K2.687.4%
    Source

    Not directly comparable

  • V*

    Grok 4.5—
    Kimi K2.696.9%
    Source

    Not directly comparable

Knowledge

  • GPQA Diamond (Vals)

    Grok 4.592.9%
    Source
    Kimi K2.689.1%
    Source

    Grok 4.5 leads this result

  • MMLU-Pro (Vals)

    Grok 4.589.2%
    Source
    Kimi K2.687.6%
    Source

    Grok 4.5 leads this result

  • GPQA

    Grok 4.5—
    Kimi K2.690.5%
    Source

    Not directly comparable

  • GPQA-D

    Grok 4.5—
    Kimi K2.690.5%
    Source

    Not directly comparable

  • HLE

    Grok 4.5—
    Kimi K2.634.7%
    Source

    Not directly comparable

Math

  • FrontierMath v2 (Tiers 1-3)

    Shared source
    Grok 4.557.193%
    Kimi K2.657.193%

    Tie

  • FrontierMath v2 (Tier 4)

    Shared source
    Grok 4.524.390%
    Kimi K2.625.641%

    Kimi K2.6 leads this result

  • AIME26

    Grok 4.5—
    Kimi K2.696.4%
    Source

    Not directly comparable

  • HMMT Feb 2026

    Grok 4.5—
    Kimi K2.692.7%
    Source

    Not directly comparable

  • MMAnswerBench

    Grok 4.5—
    Kimi K2.686.0%
    Source

    Not directly comparable

49 public results · 10 shared

Watch Grok 4.5 vs Kimi K2.6

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 7,000+ subscribers.

Last updated November 21, 2026